Skip to content

Local AI benchmarks
on hardware you’d actually buy

We test local AI with everyday use in mind, comparing task times and output quality across different hardware. Explore the results and samples to find models and configurations that help you work more efficiently. How we test.

Tested hardware

Compare hardware specs

Performance overview

Results reflect the selected setup for each system. A dash (–) indicates no reported result.

ModelBenchmark / MetricOlares OneMac StudioNVIDIA DGX SparkBeelink GTR9 Pro
Qwen3.8‑27BIFBench
Evaluation time

xhigh reasoning, 64k output cap
8.7 h12.2 h18.2 h33.1 h
Qwen3.8‑Flash‑NextIFBench
Evaluation time

300 prompts, xhigh reasoning, 65,536-token output limit
7.7 h–––
Qwen3.8‑27BGPQA Diamond
Evaluation time

xhigh reasoning, 64k output cap
11.5 h14.9 h––
Qwen3.8‑27BInference speed
Token generation, single request

20 prompts, medium reasoning
93 tok/s74 tok/s45 tok/s28 tok/s
Qwen3.8‑Flash‑NextInference speed
Token generation, single request

20 prompts, medium reasoning, 3,072-token output limit
166 tok/s130 tok/s80 tok/s–
Qwen‑Image‑2.1 BaseText to image
Time per image
12 s1.6 min27 s1.6 min
Qwen‑Image‑2.1 BaseImage edit
Time per image
17 s1.9 min37 s–
Qwen‑Image‑2.1 TurboImage edit
Time per image
5.2 s30 s13 s47 s
MiniMax H3 FastH3 v2Text to video
Sampling time

nvfp4, 720p, 5 s
78 s11 min––